首页|基于改进MKELM的红外空间锥体目标识别

基于改进MKELM的红外空间锥体目标识别

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
针对远距离探测时仅能获取目标的红外辐射强度序列、样本量有限、信噪比低而导致目标识别困难的问题，提出一种基于改进多核极限学习机(multiple kernel extreme learning machine，MKELM)的红外空间锥体目标识别方法。首先对红外辐射强度序列进行变分模态分解(variational mode decomposition，VMD)并重构，然后对重构序列进行时域特征提取，最后采用鲸鱼优化算法(whale optimization algorithm，WOA)优化MKELM的参数组合，在仿真生成的空间锥体目标红外辐射强度序列数据集上进行目标分类识别实验。实验结果验证了所提算法的有效性，同时表明所提方法具有较好的识别准确性和鲁棒性。
Infrared spatial cone-shaped target recognition based on improved MKELM
An infrared spatial cone-shaped target recognition method based on improved multiple kernel extreme learning machine(MKELM)is proposed in order to solve the problems that infrared radiation intensity sequence is the only data available at long-range detection,the sample size is limited and the signal-to-noise ratio(SNR)is usually low which lead to the difficulty of target recognition.Firstly,variational mode decomposition(VMD)and reconstruction are performed on infrared radiation intensity sequence.Then,time-domain features are extracted based on reconstructed sequences.Finally,whale optimization algorithm(WOA)is used to find the optimal combination of parameters for MKELM,and target recognition experiment is carried out on the simulated spatial cone-shaped target infrared radiation intensity sequence dataset by using improved MKELM.The experimental results verify the effectiveness,recognition accuracy and robustness of the proposed method.

infrared radiation intensity sequencespatial target recognitionvariational mode decomposition(VMD)whale optimization algorithm(WOA)multiple kernel extreme learning machine(MKELM)

王彩云、常韵、李晓飞、王佳宁、吴钇达、张慧雯

展开 >

南京航空航天大学航天学院,江苏南京 211106

北京电子工程总体研究所,北京 100854

红外辐射强度序列 空间目标识别 变分模态分解 鲸鱼优化算法 多核极限学习机

国家自然科学基金国家留学基金

61301211201906835017

2024

10.12305/j.issn.1001-506X.2024.10.02

系统工程与电子技术
中国航天科工防御技术研究院 中国宇航学会 中国系统工程学会

系统工程与电子技术

CSTPCD北大核心
影响因子：0.847
ISSN：1001-506X
年,卷(期)：2024.46(10)